obtaining the average value for a data set
조회 수: 3 (최근 30일)
이전 댓글 표시
Hello, I have a data set which contains a value for a number of stations during some years (station, date, value). Data for some days is missing. I need to create a column with annual average value per station. I appreciate your help.
station date value average(yearly per station)
1 2001/1/1 2 2
1 2001/1/3 1 2
1 2001/1/5 3 2
2 2001/1/1 2 2.5
2 2001/1/4 3 2.5
1 2002/1/2 2 1.5
1 2002/1/6 1 1.5
2 2002/1/1 2 2.5
2 2002/1/2 3 2.5 . .
댓글 수: 3
Azzi Abdelmalek
2012년 9월 4일
편집: Azzi Abdelmalek
2012년 9월 4일
can you specify the size of your data array, cell or .., how should the result looks like?
답변 (1개)
Matt Tearle
2012년 9월 4일
편집: Matt Tearle
2012년 9월 4일
It looks like the fourth column is what you want to calculate. If so, this will do the job, assuming that you have the first three columns as station, date, and value (numeric vectors except date which is a cell array of strings)
% get the years from the date strings
dv = datevec(date);
year = dv(:,1);
% get all combinations of station & year
[stlist,yrlist] = meshgrid(unique(station),unique(year));
% allocate space for the calculated average
avg = zeros(size(station));
% loop over every station/year combination
for k = 1:numel(stlist)
% find the location of this combination in the data set
idx = (year == yrlist(k)) & (station == stlist(k));
% calculate the mean
avg(idx) = mean(value(idx));
end
This could probably be simplified with prior knowledge of the station numbers and years.
EDIT TO ADD: Given your comment above, ignore the first couple of lines -- you already have year.
댓글 수: 3
Sean de Wolski
2012년 9월 6일
Why would you want to avoid the loop? The loop will likely be very fast and is easy to read/understand.
Matt Tearle
2012년 9월 6일
Possibly, although I doubt it will be neat. More importantly, why do you want to avoid loops?
참고 항목
카테고리
Help Center 및 File Exchange에서 Calendar에 대해 자세히 알아보기
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!